i haven't read your posts, but trading skew in insolation is incredibly tricky. you have to trade a lot of gross vega to make a little bit of money and as a result you have to be precise about risk factors that are not easy to observe. kelly criteria won't solve for these fundamental complexities
so, i'm not thinking of opening trades as butterflies.. i'm considering opening trades, and then converting them to risk-free butterflies if the stock moves in my direction.. not a big fan of butterflies as a stand-alone trade.. i also arrange my trades such that max loss is impossible - combining long puts with bull spreads, shit like that.. or, better stated, my max loss is always lower than the cost of the trade.. and thanks again for the help! i realized i had the kelly position sizing thing backwards after reading your replies, so tremendously helpful there!
what we're doing now is a combination of skew as mean reversion and skew as a directional assumption.. we're definitely heavily focused on maximizing number of occurrences as well.. as for the intricacies, i've definitely been flying thru those...ergodicity and non-ergodicity, fat tail distributions, and all the other little intricate correlations and considerations most people who discover skew trading never think to even look at.. the part about kelly, that's just about making sure i'm understanding the position sizing aspect correctly.. the intricacies of probabilities-based trading is a whole different topic.. in my experience tho, i must say, i've been able to do exceedingly well trading skew.. i did take some big losses in this recent downturn as there were some factors i wasn't considering prior (i.e., increased tail risk/risk of ruin, probability of max loss vs probability of max reward independent of net expected value, absorption risk, modulating the shape of your portfolio's volatility curve against the assets you're trading, etc.).. but overall the strategy has been immensely successful.. i also trade a lot lol.. i think i've done over 1,000 trades this year alone? as you said, you need a high number of occurrences..
i also follow a strategy of "buying the probability curve".. i figure if skew is essentially a mean reversion to neutral at expiry, and the "edge" you're expecting to capture is essentially the spread between the skewed and neutral volatility curves, then instead of running something like a 50/30 or 60/40 delta spread, run something like a 90/35, and essentially buy the bulk of the volatility curve for that expiration.. makes the trades far less binary, which reduces the necessary number of occurrences needed for the probabilities to pan out, also gives you a 35% probability of max profit vs only a 10% probability of max loss.. and the high delta exposure allows you to compensate for upwards IV expansion as is often seen in equities with call skew..
you’ve put on a 1000 skew positions in 100 or so trading days? who is the “we” in the first few sentences?
technically since december, so about 5 1/2 months... i say we because i have a trading friend who's been exploring the field with me, helping me map out the strategies and quantify the results... i did exceedingly well till the beginning of april, then i learned some real harsh lessons in how skewing volatility curves creates increased tail risk... i was running super narrow spreads, 60/50 delta verticals, even narrower sometimes... great when things go up, but that binary effect will destroy you when you're overexposed and have set your spreads so narrow there's no viable management strategy... i went from up 700% in about five months, to down 25%, now i'm back above breakeven and have revamped my entire portfolio strategy and distributions to account for the litany of understanding i've picked up these past 6 weeks of calculations and study...
Narrow ICs are essentially synthetic double digitals, and yeah, there isn't any trade management. They are very sensitive to vol but only bc you have to trade huge numbers. They work inside a dev but once you start to accumulate any delta position you're married to it. There isn't any dgamma but the thing is risking 10/1 so you can't afford any delta position early in the trade. You have to be price sensitive as microstructure plays a role. You can't afford to take fifteen cents in edge loss on a 50-wide SPX condor trading at 550x600. A 500 lot at $7500 in edge loss and $30-40K vols. It may take you a week to see enough decay to cover the entry, and imagine the gain on 50 flies. You can carry the same vol in a wide asym-fly with better greeks while you wait for vol to come in. People don't like to trade ATM but your wing risk is mitigated.
A buddy of mine is short 5K of the Aug (or Sep, can't recall) 3000/3100 x 4600/4700 irons and he got crushed today even though vol obv dropped. He can't DCA (no cash), can't hedge (no cash) and refused to cover. And yeah, it's a 22-23 million req.
thanks for the reply! ATM is definitely not a bad place to be esp if you're trading vol, why wouldn't you wanna capture that max extrinsic? i need to look more into SPX, just haven't been at a level where i can play in that arena yet i'm not entering narrow ICs to open tho, to clarify.. i'm trading 90/40 delta bull spreads in call skew with a 1.5:2 risk-reward ratio and 50% or better PoP... stock spikes and breaches the short call, roll it down a couple strikes and sell bear call spreads to cover the initial trade cost... selling the bear spreads makes it risk free and rolling the original short locks in a profit... let it sit till expiry cuz who cares, there's no risk on the table, just a free potential source of short delta should things fall back? that's the idea we're working with now, at least.. also playing with synthetic ratio spreads to turn tail risk zones into potential profits.. also got into "manufacturing skew" - find a relatively flat vol smile, with a small low point and a small high peak, buy the low, sell the high, you've got positive skew in a stock that doesn't "technically" have it..
So you convert the short straddle to a fly at an edge that results in a arb-fly... but you were up significantly on the combo anyway? Just cover it. I was short the bull synthetic straddle in TSLA and got cute and traded into a fly at a large credit. I was up twice the fly credit but with unbelievable terminal leverage. Guess what? TSLA projectile vomited a couple hundred points and it settled outside my wings. I have done a lot of fly conversions as I learned the nuances from a famous trader at the CBOE when I worked under my dad's JBO. The caveat is that these guys weren't holding this shit for a week--they were buying the bid and shorting the offer. Sure, flow works both ways, but you can also afford to give up some edge to short the body when you just bought 400 wings at 60 cents in edge. So my point is that there is no inherent leverage in conversion. How much vega is in a 10 lot straddle vs. a 100 lot fly?